Converting 50kW system from micro inverters to 8x SMA inverters

[peakbagger]

I think you now need a internal divider to run AC and DC in the same wireway. Its standard wireway with an add on internal separator to keep the DC wires separate from the AC. I am unsure if when you run the wires in through the side of the wireway if you have extend the conduit inside to the right compartment (which would be a PITA).

I did find this reference from 2017

Microsoft Word - 2017 NEC Code - Ch 3 Article 300 (1).docx (allstarce.com)

(1) 1000 Volts, Nominal, or Less. Conductors of ac and dc circuits, rated 1000 volts, nominal, or less, shall be permitted to occupy the same equipment wiring enclosure, cable, or raceway. All conductors shall have an insulation rating equal to at least the maximum circuit voltage applied to any conductor within the enclosure, cable, or raceway. Secondary wiring to electric-discharge lamps of 1000 volts or less, if insulated for the secondary voltage involved, shall be permitted to occupy the same luminaire, sign, or outline lighting enclosure as the branch-circuit conductors

[pclausen]

You guys must get a ton of snow up there. Here in VA we typically get 2 or 3 events each winter, and its typically only a few inches, although over the years there have been cases where we got a foot or more. I try to keep up with it and broom it off the ground mount arrays, but sometimes it gets away from me, or it falls during the night.

I got the superstrut "wall" put together and the one inverter already inside, mounted to it. Temporarily wiring of course until I get the wireway and what not installed:



So it will look something like this once the other inverters are mounted:



The gap in the top row is where the 225A sub-panel will go. They got the 6.0s and 7.0s on a pallet for me, just waiting on some more 7.7s to show up to ship the order. Hopefully this coming week.

I'm looking at this 4x4 10' long wireway:

https://www.gordonelectricsupply.com/p/Unity-44120Gflnk-4X4X120-Flt-Lid-Wwy/6175789

Would there be an issue with laying both AC and DC wires in the same wireway? That way I could get away with just having a single wireway.

[nerdralph]

I'll definitely check out the SMA Sunny Designer to see what it comes up with. I got my base info in there already, but want to double check my array tilt angles to get them just right and also enter my cable details.

It is currently showing my annual yield as 87,956 kWh where my actual with the micro inverters (and the 2 SMA 7.7s) was 67,200 kWh in 2020 and 67,535 kWh in 2021.

Keep in mind Sunny Designer doesn't factor winter snow cover. Here in Nova Scotia I find subtracting 30% from the winter production numbers roughly compensates for the snow.

[bcroe]

I would worry more about failed equipment and wiring loses, than
that level of clipping. Bruce Roe

[pclausen]

My current SMA 7.7s do clip quite a bit. Here's a screenshot from Feb 27, 2020:



The top one is a 3x12 array with 36 Talesun 270 Watt panels (9,720 watts).

The bottom one is a 2x12 + 1x8 array with 32 Talesun 275 Watt panels (8,800 watts).

I suppose the area being clipped is pretty small compared to the area under the curve, but I would think it adds up to several percent a year. Also keep in mind this was in February when the sun is pretty low in the sky.

I'll definitely check out the SMA Sunny Designer to see what it comes up with. I got my base info in there already, but want to double check my array tilt angles to get them just right and also enter my cable details.

It is currently showing my annual yield as 87,956 kWh where my actual with the micro inverters (and the 2 SMA 7.7s) was 67,200 kWh in 2020 and 67,535 kWh in 2021. So I definitely left a lot on the table from the failed inverters, associated cable loss @ 240V, and the micro inverters being rated at 250 Watts where a lot of my panels are 280 watts.

[oregon_phil]

Here's my weather station (Davis Vantage Pro2 with Solar Radiation and UV Index sensors) data for January:

https://www.wunderground.com/dashboard/pws/KVAAFTON15/graph/2022-01-29/2022-01-29/monthly

I currently have 12 panels connected to each of the 3 MPPTs on a 7700. I tried to retrieve the DC voltages from back in January, but they are long gone.

Looking at today, these were my peak values, which I know don't mean much, but they were definitely highest early this morning:



In a perfect world, I'd do 12 panels per MPPT input and with 60 panels in the pole barn array, that would work out to 5 strings of 12. But alas, Each 7.7 can only handle 7,700 watts, so I would be clipping quite a bit and I will have 2 x 7.7 inverters dedicated to the 60 panel array. The 4 x 15 would split the load perfectly between the 2 inverters.

I'll probably go ahead and do it as 4 x 15 now that its summer and then evaluate changing the configuration come December.

A couple of shots of the completed 60 panel array:

For what its worth, I have 10,220 watts of panels on 3 strings with a SMA 7.7. I have some clipping, but not much.

[nerdralph]

If irradiance is increased while the cell is kept at constant temperature, I understand that Voc will INcrease logarithmically as POA irradiance increases.

Good. And since you seem to like pedantism, it's not exactly logarithmic. You might learn something from studying electronics, as silicon PV cells are just really big diodes (PN junctions). Therefore the IV curves of a common diode like a 1N4148 and those of a PV cell are quite similar.

https://2n3904blog.com/1n4148-diode-forward-biased-i-v-curve/

As for the magnitude of the effect, with a 72-cell 440W mono PERC module, increasing irradiance from 800W/m^2 to 1kW/m^2 will increase Voc by slightly more than 3V.

[J.P.M.]

However you don't seem to understand that when keeping the temperature of a module constant, increasing irradiance, such as from 200W/m^2 to 800W/m^2, will increase Voc.

I don't think I'm agreeing with you.
I do think you have some, but not sufficient understanding of the physics of PV cells and how they operate.
If irradiance is increased while the cell is kept at constant temperature, I understand that Voc will INcrease logarithmically as POA irradiance increases.
However, such small increases will only occur when the cell is operating beyond it's MPP making that small increase even less relevant for any practical application such as the OP's.
Look at any graph of voltage vs. irradiance at constant cell temp. that also shows the locus of MPP's as f(irradiance). You'll find the locus is pretty close to vertical meaning voltage doesn't change much, if at all in any measurable way.

Also, because it is a logarithmic variation, that increase is quite small for any practical application like the OP's and is easily buried in negative coefficient of cell voltage with respect to temperature.
In fact, and in practice, the (negative) temperature coefficient of voltage, since it's a finished coefficient, includes and accounts for the effects of any logarithmic positive variation of cell voltage at constant temp. due to irradiance increases.

All that however, has little to do with either my statement that the effective radiant sky temp. can and will lower cell temps. to levels that are below levels usually calculated from common weather data and methods of calculation, or your statement that cell temps. are usually higher than ambient temp. - which I agree is usually true for most operating conditions but may not be true for near dawn conditions and so especially troublesome when used to estimate (for design) PV cell temps. for systems operating at or close to max. voltages such as what it looks like the OP may have.

If you want to talk more about the physics of a PV cell, I'd suggest you open another thread. I'm done wasting my time with this stuff.

[nerdralph]

OK, here's reality:

An energy balance in a PVcell will show that panel cell temperatures will increase with increased irradiance. That's the physics of it. Thank the First Law of Thermodynamics and the principle of conservation of energy.
Simply put, stuff gets warmer when the sun shines on it. That's called observational reality.

A cell's Voc will decrease with increased cell temperature. More reality.

Therefore, panel's Voc will decrease as the irradiance on the panel INcreases.

Your first sentence above is incorrect.

BTW, and FYI, while some consider irradiance and insolation the same thing, there are others who make definitional differences. For the purposes of this discussion, those differences, while moot, do nothing to change the incorrectness of your statement "Voc decreases with decreased irradiance".

Whether irradiance or insolation is the term used, your statement is still incorrect.

See pveducation.org, sec. 4.4, "Effect of Temperature" for an explanation of why Voc drops as irradiance (or insolation) increases.

I can't explain it any better.

Yet another red herring (straw man argument). I never disputed the relationship between Voc and temperature. I have been referring to the relationship between Voc and irradiance, independent of temperature.
"For a given temperature, Voc decreases with decreased insolation"

I've said that Voc decreases with temperature (around 0.3%/C for mono cells), and you seem to be aggressively agreeing with me. I've also said that if wind and air temperature stays the same, increased irradiance will increase cell temperature. You seem to be aggressively agreeing with me on this too.

However you don't seem to understand that when keeping the temperature of a module constant, increasing irradiance, such as from 200W/m^2 to 800W/m^2, will increase Voc.

[J.P.M.]

The IV curves and cell performance data prove Voc decreases with decreased irradiance (I was writing to quickly and wrote insolation by mistake).
Even without looking at datasheets, if cell temperature stay the same, it should be obvious that Voc will be lower at night than during the day in bright sun (>800W/m^2 if you want a specific definition).

OK, here's reality:

An energy balance in a PVcell will show that panel cell temperatures will increase with increased irradiance. That's the physics of it. Thank the First Law of Thermodynamics and the principle of conservation of energy.
Simply put, stuff gets warmer when the sun shines on it. That's called observational reality.

A cell's Voc will decrease with increased cell temperature. More reality.

Therefore, panel's Voc will decrease as the irradiance on the panel INcreases.

Your first sentence above is incorrect.

BTW, and FYI, while some consider irradiance and insolation the same thing, there are others who make definitional differences. For the purposes of this discussion, those differences, while moot, do nothing to change the incorrectness of your statement "Voc decreases with decreased irradiance".

Whether irradiance or insolation is the term used, your statement is still incorrect.

See pveducation.org, sec. 4.4, "Effect of Temperature" for an explanation of why Voc drops as irradiance (or insolation) increases.

I can't explain it any better.

[bcroe]

A couple of shots of the completed 60 panel array

On the ground array, your use of it as a shelter would conflict with my idea of
putting vertical snow gaps between rows of panels. Bruce Roe

[nerdralph]

In a perfect world, I'd do 12 panels per MPPT input and with 60 panels in the pole barn array, that would work out to 5 strings of 12. But alas, Each 7.7 can only handle 7,700 watts, so I would be clipping quite a bit and I will have 2 x 7.7 inverters dedicated to the 60 panel array. The 4 x 15 would split the load perfectly between the 2 inverters.

One thing I like about SMA's Sunny Designer is that it estimates clipping. Usually you can do a 1.2:1 DC:AC ratio with clipping less than 0.5% of annual production.

[nerdralph]

BTW, your statement "Voc decreases with decreased insolation" is incorrect.

There's no point in debating if you demand I provide proof/evidence of all my points and you provide none. My points are easily verifiable with public data, such as module datasheets. The IV curves and cell performance data prove Voc decreases with decreased irradiance (I was writing to quickly and wrote insolation by mistake).

https://www.canadiansolar.com/wp-content/uploads/2019/12/Canadian_Solar-Datasheet-BiHiKu_CS3W-MB-AG_EN.pdf

Even without looking at datasheets, if cell temperature stay the same, it should be obvious that Voc will be lower at night than during the day in bright sun (>800W/m^2 if you want a specific definition).

[J.P.M.]

That's a red herring. I specifically said "in bright sun", not at the break of dawn. For a given temperature, Voc decreases with decreased insolation.

Well, I've provided an example of one of many empirical correlations in the literature used to estimate array cell temps. The one I suggested is one of many I've verified and found it to be reasonably close to what I've measured several hundred times.

The flaw in your statement is "bright sun". First off What is bright sun ? At what time ? Or more specifically, at what angle of beam irradiance on the array ? and what irradiance level ? Your statement is too general. That's the real red herring. Your statement is, IMO only, vague to the point of being useless.

You also mentioned "on a cold winter day". The reality that makes that statement irrelevant is that the ambient air temp. by itself has little to do with the ambient air to cell temp. difference.
That temp. difference is essentially independent of the ambient air temp. I'd explain it to you but space is limited here. Consult a text on convective heat transfer if you're curious. I've got several titles on by bookshelves I can recommend.

Here's reality and Physics: It doesn't matter if the air temp. is +40 C or -40 C, or any other temp. Under what are otherwise the same environmental conditions of wind vector, irradiance and sky cover conditions, the temp. differences, ambient air to cell will be essentially the same regardless of air temp. There can be very small differences due to what are called the transport properties of (in this case) air as f(air temp.), but those differences are very small and usually ignored.

But most of that has little to do with my purpose in providing information to the OP about effective radiant sky temperature.
Since the OP is (or may be) operating at close to maximum voltages, I mentioned that the array temp. can be lower than the ambient air temp. on clear, mostly windless, low dew point mornings shortly after sunup due to low effective radiant sky temps. only for informational purposes and to suggest that array temps. can be lower than a lot of correlations and rules of thumb (and stature) state which might indicate a design change for voltage considerations, especially when operating at close to max. voltages.

Also, I believe your statements - which you seem to stand by with no corroborating information or logic - only accusation - are sloppy and unsubstantiated, and may lead the OP to problems if he's operating close to inverter max. voltages.

If you have some meaningful information to back up your hip shooting that makes sense, I'd be happy to read it. Otherwise, I'd suggest being considerate of others and stick to what you know.

BTW, your statement "Voc decreases with decreased insolation" is incorrect.

[pclausen]

Here's my weather station (Davis Vantage Pro2 with Solar Radiation and UV Index sensors) data for January:

https://www.wunderground.com/dashboard/pws/KVAAFTON15/graph/2022-01-29/2022-01-29/monthly

I currently have 12 panels connected to each of the 3 MPPTs on a 7700. I tried to retrieve the DC voltages from back in January, but they are long gone.

Looking at today, these were my peak values, which I know don't mean much, but they were definitely highest early this morning:



In a perfect world, I'd do 12 panels per MPPT input and with 60 panels in the pole barn array, that would work out to 5 strings of 12. But alas, Each 7.7 can only handle 7,700 watts, so I would be clipping quite a bit and I will have 2 x 7.7 inverters dedicated to the 60 panel array. The 4 x 15 would split the load perfectly between the 2 inverters.

I'll probably go ahead and do it as 4 x 15 now that its summer and then evaluate changing the configuration come December.

A couple of shots of the completed 60 panel array: